[Geoserver-devel] Re: [Geotools-devel] GeoAPI 2.0 and Geotools 2.1

Martin Desruisseaux wrote:

Jody Garnett a écrit :

I am not sure much should be done about this until such time as geotools goes to 1.5, I certaintly find working with collections to be easier.
Fortunantly in udig we often make use of code like the following:

for ( CoordinateReferenceSystem crs : compound.getCoordinateReferenceSystem() ){
       //
}
So when you do get a chance to go to 1.5 collections a fair bit of code should not break.

But we may not be able to change the type after GeoAPI release. So the conclusion may be: live with arrays for those methods - definitively.

The other way to work is:
- issue a date stamped GeoAPI / Geotools 2.1 / uDig 1.0 / GeoServer 1.3
- and after your change GeoAPI 2.0 / Geotools 2.2 / uDig 1.1 / GeoServer 1.4

Rather then fuss about *what ifs* can you email me a revised geoapi jar, I can drop
it in my local udig environment and we can estimate how long it would take to update udig.
You should probably ask the same question of David Blasby.

I understand and respect that it would be nice to work with a stable api right across the board,
I still have my concerns about the reduction in expressiveness of the api change, but I like collections
enough to live with the occasional cast, the warnings produced by not typed collections use in a
Java 5 program like udig are going to drive me a bit batty.

Once again if you can send me the new geoapi jar I can drop it in and see how many udig compile
errors result.
Jody

Jody Garnett a écrit :

The other way to work is:
- issue a date stamped GeoAPI / Geotools 2.1 / uDig 1.0 / GeoServer 1.3
- and after your change GeoAPI 2.0 / Geotools 2.2 / uDig 1.1 / GeoServer 1.4

It work for peoples who know each other. May hope was to get a really stable release if we can.

Rather then fuss about *what ifs* can you email me a revised geoapi jar
I can drop it in my local udig environment and we can estimate how long it would take to update udig.

Done by private email, for avoiding to pollute the list with attached files. If someone else would like to test the JAR, please let met know.

You should probably ask the same question of David Blasby.

If we wants (maybe after yours result? If the answer is "no" from uDig side, it may not be worth to bother peoples on the GeoServer side).

  Martin.